[Zoobank-list] RE: ZooBank Data Objects
Richard Pyle
deepreef at bishopmuseum.org
Wed Mar 7 02:47:50 GMT 2007
> You said:
> >I guess the way I would define a "usage instance" would be the
> >treatment of one "name" in one "documentation instance". A
> publication
> >would be the most familiar form of "documentation instance", but I
> >prefer to keep it broad and generalized.
>
> as I said, parallel, but not equal. A "usage instance" could
> also include a specimen determination, among other things.
Yes, perhaps (or maybe specimen determinations should be treated as a
separate class) -- but fundamentally it's the same thing. Publications are
just a subset of "documentation instances", as far as I can tell.
> You said:
> > If, in a publication,
> >the name "Aus bus" is used consistently four times on page 4, and 2
> >times on page 8 and again in a table on page 9, I would not think of
> >that as 7 discrete usage instances; but rather as one usage
> instance,
> >representing the treatment of a name within the context of a
> specific
> >unit of documentation instance (e.g., a publication).
>
> I could not agree more. In fact, it has been my intention to
> make sure that all those 'mentions' (if I may use that term)
> in the same work are linked to a single GUID.
Perfect! I like "Mentions". I'm tempted to use "Appearance", but I don't
want to confuse it with James Ytow's notion of an "Appearance".
> I have discussed examples similar to this one with a number
> of zoologists over many years. We disagree on the term (Zoo:
> senior/junior synonyms = Bot. taxonomic synonyms).
True.
> We also
> (botanists and zoologists) have different definitions of
> nomenclatural acts.
Very true! But the question is, can we converege on what constsitutes a
"usage instance" (by whatever name) -- one subset of which constsitute
nomenclatural acts in zoology, and another subset of which constitute
nomenclatural acts in botany.
> Botanists would say that Brown and
> Weitzman had made new combinations, which are nomenclatural
> acts.
Yes, agreed -- and that's the heart of the difference between the meaning of
a "name" in botany vs. zoology. But I think we can easily accommodate both
needs via usage instances.
> However, for zoologists, these are not nomenclatural
> acts--however, they are still combinations (or forms of a
> name) that you use in the literature. For that reason, I
> would argue that the place where these combinations are first
> made is worth recording (if for no other reason than that
> someone doesn't have to search for where it was done the
> first time each time someone works on a species--and we are,
> after all, trying to speed up the rate that taxonomy can be
> done by creating these databases and linking GUIDs, etc.).
I agree completely -- even though these "first use of a combination"
instances are not governed by the ICZN Code, they are of value to zoological
taxonomists, and hence should fall within the scope of ZooBank (in my
opinion). But that's not the issue, the issue is should they be given new
"name" GUIDs (if such things exist)? A botanist would say "of course", and
a zoologist would say "of course not". I think we can avoid the problem
entirely by having both camps view them as usage instances, and having the
botanical camp deem them to be usages that constitute nomenclaural acts, and
zoological camp deem them to be usages of interest, but otherwise not
governed by the Code. The important thing would be that both botanists and
zoolgists treat them as usage instances, in accordance with a shared
definition and data structure behind that term.
> Most zoologists I have discussed this with have agreed with
> me that recording the 'homotypic' synonyms and where they
> were first made is useful and convenient whether or not they
> have nomenclatural standing.
This zoologist would agree as well -- but I see this as a separate issue
from whether usage instances are always directly "anchored" to "basionyms".
> You said that in this case, the later combinations would be
> 'pointers' to the basionyms. I think that is where we
> diverge. As a botanist, I would give all four a GUID, but
> each of the homotypic synonyms should have 'pointers' to
> their basionym as well (within the work and within a
> registry), and in this particular work the 3 would also have
> pointers to the 'accepted name'='senior synonym' (although
> most botanists would only see 3 synonyms and not
> differentiate them, except implicitly in their minds).
I'm not sure I follow, so maybe we need to be explicit with examples. Yes,
certainly, there need to be GUIDs for the different combinations, but the
question is do we need a GUID for the usage of each combination *within the
context of Jones* (in the example I gave).
So, imagine we have these five publications, each represented by a GUID, but
for the sake of clarity I'll represent those GUIDs as follows:
Linnaeus-1753
Smith-1850
Brown-1976
Weitzman-2000
Jones-2006
Each of these publications might have among their name usage instances the
following (denoted by "UI" GUID numbers):
Publication: Linnaeus-1753
GUID BasGUID Rank Spelling ParentGUID
ValidGUID
---------------------------------------------------------------------
UI1 UI1 Genus Aus -
UI1
UI2 UI2 Species Aus xus UI1 UI2
Again, "UI" numbers are Usage instance GUIDs.
"BasGUID" is a pointer to the "first creation" of a name (loosely the
"basionym" -- but as I mentioned before, this is not the best term). In
both cases above, these names were first created within *this* publication
(Linnaeus-1753), which is why the BasGUID points back to itself, indicating
that both of these usages instances *are* the original creation of the name.
So, we have two usage instances within the publication Linnaeus-1753, and it
just so happens that both of these represent the first creations of a new
"basionym" (as I said, not the best term).
===============================
Publication: Smith-1850
GUID BasGUID Rank Spelling ParentGUID
ValidGUID
---------------------------------------------------------------------
UI3 UI1 Genus Aus -
UI3
UI4 UI4 Species Aus pylei UI3 UI4
In this case, Smith-1850 referred to the genus name "Aus" as created by
Linneaus-1753 (BasGUID=UI1), and treated it as valid (GUID=ValidGUID).
Smith-1850 also created the new species epithet "pylei" (GUID=BasGUID), and
placed it within his treatment of the genus Aus (ParentGUID=UI3).
===============================
Publication: Brown-1976
GUID BasGUID Rank Spelling ParentGUID
ValidGUID
---------------------------------------------------------------------
UI5 UI5 Genus Anothergenus - UI5
UI6 UI2 Species Anothergenus xus UI5 UI6
This is similar to the Smith-1850 case, except in this case, Brown-1976
created a new genus name "Anothergenus" (GUID=BasGUID) [UI5], and
established a new combination of "Anothergenus xus" [UI6].
===============================
Publication: Weitzman-2000
GUID BasGUID Rank Spelling ParentGUID
ValidGUID
---------------------------------------------------------------------
UI7 UI5 Genus Anothergenus - UI7
UI8 UI4 Species Anothergenus pylei UI7 UI8
In this case, no new "Base" ("basionym") names were created, but there are
two usage instances, one [UI7] for "Anothergenus" (BasGUID=UI5), and the
other [UI8] that establishes the new combination "Anothergenus pylei" for
the basionym "Aus pylei Smith" (BasGUID=UI4).
Finally:
===============================
Publication: Jones-2006
GUID BasGUID Rank Spelling ParentGUID
ValidGUID
---------------------------------------------------------------------
UI9 UI1 Genus Aus -
UI9
UI10 UI2 Species Aus xus UI9 UI10
UI11 UI4 Species Aus pylei - UI10
In this case, we have no new "Base" names and no new combinations -- just 3
plain old usage instances without any nomenclatural standing in either Code.
The only think new here is that for UI11, the name was treated as a
taxonomic synonym of UI10 (as indicated by ValidGUID=UI10, for the record of
UI11). I've indicated the "ParentGUID" of UI11 as empty ("-"), because
technically if the name is not regarded as valid, it inherits the aprent of
the name that it is a taxonomic synonym of. But in any case, from a taxon
concept perspective, it would be easy enough to subsume UI11 under the
"umbrella" of UI10, so that one GUID (UI10) represents the entire taxon
concept. But that's a topic for another day....
Up until this point, I think the needs of both botany and zoology are
equally well served. We have eleven usage instances, each identified by a
GUID (UI1-UI11). Each represents a single treatment of a single "name"
within a single publication. Of these eleven usage instances, four (UI1,
UI2, UI4 & UI5) would represent Code-governed nomenclatural acts under both
ICZN and ICBN. Two others (UI6 and UI8) represent nomenclatural acts under
ICBN, but not under ICZN. However, as you point out, even though these two
usage instances are not Code-governed nomenclatural under ICZN, they are
certainly of interest to zoologists, and would very likely fall within the
scope of ZooBank. None of the last three usage instances (UI9, UI10, UI11)
constitute nomenclatural acts under either Code, but users in both realms
would find the information equally useful. The same could be said for UI3
and UI7.
So far, we've captured the useful data in a way that satisfies the needs of
both nomenclatural domains.
Now...here's the sticky part. How do we deal with the appearance
("mention") of "Anothergenus xus (L.) Brown" and "Anothergenus pylei (Smith)
Weitzman" within the context of the publication "Jones-2006". Both of these
"names" (sensu Botany) have GUIDs associated with them (UI6 and UI8,
respectively). So from a botanical perspective, these "names" have
representative GUIDs. And, we can develop a data structure that tracks the
fact that the publication "Jones-2006" acknowledge the existience of both
UI6 and UI8, without having to create new usage-instance GUIDs within the
context of the Jones-2006 publication.
We certainly could treat them as new usage instances (with GUID's) along the
lines of the following:
Publication: Jones-2006
GUID BasGUID Rank Spelling ParentGUID
ValidGUID
---------------------------------------------------------------------
UI12 UI6 Species Anothergenus xus - UI10
UI13 UI8 Species Anothergenus pylei - UI10
...but I don't understand what you would gain by doing so. We already know
from UI6 and UI10 that they share a common basionym (UI2); so establishing a
new GUID-record for UI12 doesn't tell us anything new, really. Same
situation with U13. In zoology, we certainly would want to create separate
GUIDs for UI6 and UI8, even though we don't treat these as Code-governed
acts. But it would be utterly nonsensical from a zoologist's perspective to
create separate GUIDs for UI12 & UI13, because the important information is
already documented via UI10 and UI11. With UI12 and UI13, the "BasGUID"
links do not go directly to "original name creations" -- it requires a
second hop to get to UI2 and UI4 (respectively).
>From a taxon concept perspective -- yes, for sure, we want to know what
Jones thought of the treatement of "Anothergenus pylei (Smith) Weitzman SEC
Weitzman-2000" [UI8], as compared to "Aus pylei Smith SEC Smith-1850" [UI2],
or any other concept definition involving the epithet "xus Smith, 1850", but
we've got GUIDs already to reprsent each of those.
But even if the botanical world decides it needs to define "usage instances"
to include the likes of UI12 and UI13, I still think we're much better off
treating all of these as "usage instances", rather than "name objects".
Then the role of ZooBank and IPNI and IF and such becomes one of identifying
and managing special-case usage instances that fall within the governances
of their respective Codes (no matter that the scope of govenrance may
differ), but the point is that we are at least converging on a common system
of GUIDs that we can share universally without conflict.
> >We certainly may want to record that
> >fact, of course (which is fine), but my point is that the linking of
> >?"Anothergenus pylei (Smith) Weitzman" as a junior synonym
> of "Aus xus L."
> >tells us nothing nomenclaturally about Jones' treatments
> that we don't
> >already know by the listing of "Aus pylei Smith" as a junior
> synonym of
> >"Aus xus L." Having a separate index to cross-reference Jones'
> >usage-instance GUID for "Aus xus L." to Weitzman's usage
> instance GUID
> >for "Anothergenus pylei (Smith) Weitzman" is great -- but I
> don't see a
> >need to create yet another usage instance GUID in the
> context of Jones'
> >publication specifically for this link. I see that more as metadata.
>
> This is an interesting point, and I see your point since in a
> functioning database it would not matter, since all of the
> homotypic synonyms would move together no matter what we did.
Yes!
> However, I fear that it will be a long time before most
> botanists will accept this.
But here's the thing: most botanists don't *need* to accept it -- only
botanical database developers would need to embrace it. Because the power
of a solid database structure is that you can render the information back to
the user any way you want. 99% of users will not know what's going on
behind the scenes at the database structure level -- they only care about
what they see on their computer screens. So as long as the data structure
can produce for them what they want to see, everyone is happy.
I have to be careful here not to present the zoological approach as "better"
-- because it's not. We're only talking about conventions here. In truth,
the actual underlying data works in exactly the same way for both botany and
zoology. Sure, we have different rules about what is allowable and what
constitutes a nomenclatural act (e.g., we in zoology don't accept any
infraspecific ranks except subspecies). But at the data level, it's all the
same: a hierarchical structure of names, cross-linked to publications.
Below the level of genus, we always render those names with the full
combination, to conform to the practice of binomial nomenclature (no
problems there). Subgenera are treated similarly for both Codes. And
really, most of the rules that govern how names become available are
fundamentally similar. My motivation in these impassioned pleas is to let
the structure of the data guide us in defining GUIDs and protocols of data
exchange. Whether you regard "Anothergenus pylei (Smith) Weitzman" as a
separate "name" from "Aus pylei Smith" doesn't really matter in terms of
data structure -- that's just an issue of how you present the information to
the user. As far as the data structure is concerned, what we want to know
is that Smith originally created the new species epithet "pylei", and
originally placed it in the genus "Aus" (which itself was originally created
by Linnaeus), and that Weitzman was the first to place Smith's epithet "xus"
in combination with Brown's genus "Anothergenus", and that Jones regarded
the type specimen of Smith's epithet "pylei" as being conspecific with the
type specimen of Linnaeus' species epithet "xus". These are the facts
around which the data structure and exchange protocols should be developed.
I should also point out something else: even if Botany and Zoology had
exactly the same notion of what a "name" was -- or if both codes were
already unified under a common "Biocode" -- I would *STILL* be just as
impassioned about breaking away from the idea of assigning GUIDs to "names",
rather than to usage instances. I've learned over the years that the
"natural" structure of information is not always in perfect synch with the
way that information is traditionally used by its consumers. Every time
I've gone the "easy" route to structure the data in a way that the users
expect to see it, instead of structuring it the way it naturally exists,
I've ended up regreting it in the long term. If you structure the data the
way it "naturally" exists, you can always render it back to the form the
users want to see. But the reverse is seldom true.
> Though perhaps if we could agree
> that there is a different kind of linkage between heterotypic
> and homotypic synonyms--which is not a usage instance link or
> GUID we could all be happy :-) ?
I think that would certainly be a step in the right direction! Also, there
are different kinds of "homotypic" synonyms. In most cases, they would be
"homobasionym" synonyms (i.e., Aus xus Linnaeus = Anothergenus xus
(Linnaeus) Brown]), but there are other cases (e.g., replacement names in
zoology, which I believe are the same as "Nom. Nov." in botany). In rare
cases, the same specimen has been used as the holotype for more than one
species epithet. And there are other ways that names are homotypically
synonymous without stemming from the same "basionym".
> You make an interesting point about the word "basionym". I
> have struggled for years to find a widely accepted
> definition--or one sanctioned by the botanical Code. It has
> been used both in the way that you use it, and for database
> purposes at least, also in place of your protonym. However,
> it is only the new Code (published last fall) that finally
> includes a glossary. In that glossary the word is finally defined:
>
> "basionym. A previously published legitimate name-bringing or
> epithet-bringing synonym from which a new name is formed for
> a taxon of different rank or position (Art. 33.4, 49.1 and 52.3)."
I just remember this discussion:
http://wiki.tdwg.org/twiki/bin/view/UBIF/LCProtonymDiscussion
"a name can *have* a basionym but can't *be* a basionym"
> This definition does indeed read the way that you have used
> it and excludes 'a previously published legitimate name' that
> has not been used to form a new name at a different rank or
> position. So I guess we will have to adopt your term protonym.
I've been carefully trying to avoid using that term "Protonym", both because
of F. Bungartz's comments here:
http://wiki.tdwg.org/twiki/bin/view/UBIF/LinneanCoreDefinitions#ProtonymDefi
nition
...and because I didn't want to turn this into a sales pitch for the
Taxonomer approach to structuring nomenclatural data.
But of course, I would be happy to adopt that term.
Anyway, as is probably apparent, I find these discussions highly interesting
*and* productive. Apologies to those who would disagree.
Aloha,
Rich
Richard L. Pyle, PhD
Database Coordinator for Natural Sciences
and Associate Zoologist in Ichthyology
Department of Natural Sciences, Bishop Museum
1525 Bernice St., Honolulu, HI 96817
Ph: (808)848-4115, Fax: (808)847-8252
email: deepreef at bishopmuseum.org
http://hbs.bishopmuseum.org/staff/pylerichard.html
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
More information about the Zoobank-list
mailing list